reviewer

Review candidate diffs for scope violations, missing tests, and runtime equivalence.

Updated Apr 15, 2026
One-click install
npx skills add https://github.com/kyleyhw/agent-evolve --skill reviewer-kyleyhw
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: reviewer
Source: https://github.com/kyleyhw/agent-evolve/tree/main/.claude/skills/reviewer
Command: npx skills add https://github.com/kyleyhw/agent-evolve --skill reviewer-kyleyhw

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

The reviewer acts as the final gatekeeper to protect the codebase by enforcing critical checks on every candidate. It evaluates scope compliance, ensures test/metric improvements are preserved, validates runtime equivalence in the presence of hypothesized changes, and assesses the coherence of the explorer's hypothesis and conclusion before a candidate can be merged or promoted to a final pull request.

Core Features & Use Cases

  • Gate candidates by scope: reject diffs that touch do-not-touch areas or exceed target_files.
  • Validate test coverage and metric improvements to prevent regressions.
  • Verify runtime-mode equivalence and coherence between explorer hypotheses and conclusions.
  • Emit structured verdicts (APPROVE / REQUEST_CHANGES / REJECT) with a detailed checklist for human review.

Quick Start

Review the candidate diff and issue a verdict with a structured checklist.

Frequently Asked Questions about reviewer

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I enforce scope violations and missing test checks on a pull request?

To enforce scope violations and missing test checks on a pull request, the reviewer evaluates candidate diffs against target_files and do-not-touch areas, rejecting non-compliant changes. It provides a structured verdict and detailed checklist to prevent regressions and ensure code quality.

What is runtime equivalence validation when reviewing code diffs?

Runtime equivalence validation in code review verifies that hypothesized changes do not alter expected runtime behavior. The reviewer assesses the coherence between the explorer's hypothesis and conclusion to ensure no unsafe changes or regressions slip into the main branch.

How do I generate a structured quality assurance verdict for a code candidate?

You generate a structured quality assurance verdict by reviewing the candidate diff to issue an APPROVE, REQUEST_CHANGES, or REJECT decision. The reviewer outputs a concise verdict alongside a detailed checklist and confidence assessment for human review.

Can I use automated gatekeeping to reject diffs that exceed target files?

Yes, you can use automated gatekeeping to reject diffs that exceed target files. The reviewer acts as a final gatekeeper, identifying scope compliance violations and blocking candidates that touch do-not-touch areas before they can be merged or promoted to a pull request.

When do I need a final gatekeeper to assess code coherence before merging?

You need a final gatekeeper to assess code coherence before merging when validating runtime-mode equivalence and explorer hypothesis alignment. The reviewer ensures test and metric improvements are preserved and validates that no regressions or unsafe changes occur in the codebase.

What are the limitations of relying on automated code diff reviews for quality assurance?

Automated code diff reviews for quality assurance rely heavily on the accuracy of the explorer's inputs and defined scope boundaries. The reviewer assesses coherence and reports scope violations based on these inputs, meaning incomplete hypotheses or misconfigured target_files can limit its gatekeeping effectiveness.